CODEWITHSUNDEEP
pythontensorflowmachine-learningdeep-learningtensorflow-lite
이지연
이지연

Reputation: 49

How can I know Conv2D parameters inside tflite model using python?

I was able to find out each Conv2D's input/output tensor shape inside tflite model with below code.

import tensorflow as tf

SAVED_MODEL_PATH = "TFLITEMODEL_PATH.tflite"
interpreter = tf.lite.Interpreter(model_path=SAVED_MODEL_PATH)

ops = interpreter._get_ops_details()
for op_index, op in enumerate(ops):
    if op['op_name'] == "CONV_2D":
        cnt += 1
        for tensor_idx in op['inputs']:
            tensor = interpreter2._get_tensor_details(tensor_idx)
            tensor_shape = tensor['shape']
            print(tensor['name'], "\t", tensor['shape'])
        print("----")

And Below is the output.

Placeholder      [  1 224 224   3]
conv2d/kernel    [64  7  7  3]
conv2d/Conv2D_bias   [64]
----
block-0/denseblock-0-0/Relu      [ 1 56 56 64]
block-0/denseblock-0-0/conv2d/kernel     [32  3  3 64]
block-0/denseblock-0-0/conv2d/Conv2D_bias    [32]
----
block-0/denseblock-0-1/Relu      [ 1 56 56 96]
block-0/denseblock-0-1/conv2d/kernel     [32  3  3 96]
block-0/denseblock-0-1/conv2d/Conv2D_bias    [32]
----

But I wonder how can I know its Conv2D parameters(like padding, stride, dilation, etc) with python code. I want to those information like netron.app. It shows all layers and its info like name, padding, stride, etc. enter image description here

Upvotes: 0

Views: 675

Answers (1)

miaout17
miaout17

Reputation: 4875

There is no official way to do that. _get_ops_details isn't a public API and isn't guaranteed to be stable.

May I know what you're trying to achieve?

Technically it's possible to go into the detail, and parse the TFLite FlatBuffer model by your own. However it's not a official path either.

Upvotes: 1

Related Questions